There are known permanent magnet structures, notably those of the so called plate-magnet type. Structures such as these are described, for example, in the European patent application No. EP-A-0 170 318. Or, again, structures such as these are described in the U.S. Pat. Nos. 4,672,346; 4,679,022 and 4,818,966. Other types of permanent magnet structures are also already known: these are notably a spherical permanent magnet with equatorial access as described in the French patent application No. 2 605 452.
The drawback of the so-called plate-magnet structures lies in the closing of the magnetic field lines. Indeed, to put it in the simple way, in a plate-magnet structure, the transverse field is produced in an air gap located between two plates of permanently magnetized material. The southern face of one plate faces the northern face of the other plate on either side of the gap. The closing of the magnetic field lines is therefore organized between the faces that are furthest away from the two plates. To avoid detrimental effects on the value of the field produced in the air gap by the existence of this other air gap for the closing of the field lines, it is the usual practice to position, between these two distant faces, closing structures made of a magnetizable magnetic material, typically soft iron. These soft iron structures have several drawbacks. Firstly, although the cost of soft iron is low, it nevertheless adds to the price of the magnet. Furthermore, these soft iron structures take up space and limit the points of access to the useful zone of the air. The ideal solution would be to close the field lines everywhere. However, in this case, it is no longer possible to enter the zone of interest of the magnet.
Furthermore, the approaches thus conceived firstly do not lead to great homogeneity of the intense magnetic field in the zone of interest and, secondly, are most likely to give rise to eddy currents during the sequences for the excitation and measurement of the NMR signal used in the imaging methods. The homogeneity is not high for it can only be empirical. Indeed, these soft iron structures, which are also used to hold the magnetic plates, have shapes dictated by considerations of mechanics that do not lend themselves to the preparation of precise models of their contribution to the magnetic field. It is possible to use only techniques of computation by finite elements which do not have sufficient precision with respect to the homogeneity of the field required by the NMR.
The fact remains, therefore, that it is necessary to carry out the empirical validation of the actually created field. This empirical validation makes it necessary to conceive of the positioning of soft iron parts at certain particularly appropriate places so as to measure the effects of this positioning and carry out the consequent modification until, little by little, the requisite modification is achieved. This empirical technique is not an industrial-scale technique: great skills are required to make only one copy, and the technique is not easily reproducible.
With regard to eddy currents, the pulsed characteristics of the gradients of the additional magnetic field, applied during the sequences for the excitation and measurement of the NMR signal in the examinations, run counter to such an approach because soft iron is conductive. In the invention, as shall be seen further below, a structure without soft iron is used, implementing solely blocks of magnets which, furthermore, can be computed analytically. If these blocks of magnets are ferrite-based blocks, they are themselves insulators and are therefore not the site of eddy currents.
In the French patent application relating to the spherical permanent magnet with equatorial access referred to here above, a symmetrical structure was made, based on rings with a generally hemispherical shape producing a transversal field. The drawback of this structure lay in the fact that, one the one hand, the holding and adjusting of the spherical rings with respect to one another is not a simple matter in mechanical terms. And, above all, secondly, the zone of interest in which the field is homogeneous is far greater than the dimension of the point of access to this zone. In other words, the two hollow hemispherical structures give a large zone of interest with great homogeneity in their interior.
By contrast, the access to the interior is narrow; it is as if it were assumed that the spacing between these structures were not enough to enable sliding into them. In practice, the medical machines made with this invention are ill-suited to the examination of adult patients. Furthermore, the magnetization of the magnetic blocks used to make the rings of this spherical structure had to have an orientation as a function of the position of these blocks. This position was referenced on a circle having one of its diameters perpendicular to the plane of the equatorial access to this magnet. This was also true for the value of magnetization of these blocks. This led to difficulties in the making of the blocks.